Klasifikasi Dokumen Bahasa Indonesia Menggunakan Adaptive Classifier Combination (ACC)

Abstract

Search engine is a tool to find an information that match with user’s request based on query or keywords. However, the query of search engine often shows irrelevance sites that didn’t match with user’s desire. Classification was developed to focused user’s request on the information that they searched, so the result would be more accurate and less effort yet effective and efficient for the system. Adaptive Classifier Combination is one of many methods that can be used for text classification. This method is a combination between k-nearest neighbor classification and Naive Bayes classification to get a better result than using single method. The result of this research is an implementation of web-based application that developed using PHP language. The document is from horticulture reasearch journal from 2002 until 2009. From 340 documents, 300 document is used for data training and 40 documents is for testing. The accuracy of ACC is 80% and for Naive Bayes classification is 42.5%.